Pitaka 0.6.0

Added

• Scroll the reader's text with the up and down arrow keys, even after
  clicking a chapter in the sidebar. Holding a key repeats, and the
  shortcuts sheet lists them.

Fixed

• The "Bookmarks & notes" menu item showed as "Bookmarks _ notes"; the ampersand now displays.

Downloads: Windows: Pitaka_0.6.0_x64-setup.exe. Linux: the .AppImage, or the .deb/.rpm for your distro. The Windows installers aren't code-signed yet, so Windows SmartScreen will warn before running them; choose More info → Run anyway. SHA256SUMS lists each file's checksum.

Pitaka 0.5.0

Added

• Covers: each book's cover is taken from its EPUB (or a PDF's first
  page) and shown on the Books screen and Home. The Books screen has a
  Grid view of cover cards; books without a cover get a generated one.
• Reopen each book where you left it. A progress bar and "Chapter 2 of
  3 · 52%" show how far through a book you are, and Home has a Continue
  reading section.
• Reader settings: a Display button sets text size, line spacing,
  column width and a System, Light, Sepia or Dark theme. The text is
  also set better, with old-style figures, a heading for the chapter
  title and small caps on the first line.
• Back up the library to one .pitaka file and restore it, with the
  previous library kept as an undo.
• Edit a book's title and author from the Books screen.
• Search within a single book, from the Search screen's "Search in"
  picker or the reader's Search this book button.
• Filter the Books screen and sort it by title.
• Keyboard shortcuts (press ? for the list) and a File / View / Help
  menu. The window opens larger, and remembers its size and position.
• Progress, loading and empty states across the app, with icons and
  motion that respects reduced-motion settings.
• Home and the Help menu link to bug reports, ideas and email feedback.

Changed

• Text, spacing and buttons use one consistent scale, and buttons have
  a clear primary, secondary, quiet and danger style.

Fixed

• Words that OCR'd PDFs break with soft hyphens are rejoined, and a
  paragraph that spans three pages is no longer split in two.

Downloads: Windows: Pitaka_0.5.0_x64-setup.exe. Linux: the .AppImage, or the .deb/.rpm for your distro. The Windows installers aren't code-signed yet, so Windows SmartScreen will warn before running them; choose More info → Run anyway. SHA256SUMS lists each file's checksum.

Pitaka 0.4.0

Added

• Import PDFs: "Import book…" and "Import folder…" now take PDFs that
  have a text layer, alongside EPUBs. Chapters come from the PDF's
  outline (or are 20-page blocks without one), running headers and page
  numbers are dropped, and words hyphenated across lines are rejoined.
  The title comes from the PDF's metadata, else its title page. Scanned
  PDFs and ones that need a password to open are refused with a message
  saying why.
• Index books already in the library for chapter search: the Books
  screen has an Index button on each book that isn't fully indexed,
  Index all books, and Stop indexing. Books are indexed one at a time
  in the order queued, and a stopped run carries on from the last
  finished chapter.

Changed

• "Import EPUB…" is now "Import book…", and the folder import's Stop is
  now Stop import.

Downloads: Windows: Pitaka_0.4.0_x64-setup.exe. Linux: the .AppImage, or the .deb/.rpm for your distro. The Windows installers aren't code-signed yet, so Windows SmartScreen will warn before running them; choose More info → Run anyway. SHA256SUMS lists each file's checksum.

Pitaka 0.3.0

Added

• Installers for Windows (-setup.exe and .msi) and Linux
  (.AppImage, .deb and .rpm) on each GitHub release, with chapter
  search by meaning built in. The Windows installers aren't code-signed
  yet, so SmartScreen warns before running them.
• Highlight and write notes while reading: select text in a paragraph
  to highlight it in one of four colours, and add a note to a highlight
  or a whole paragraph. Notes open in the margin beside the text. The
  Bookmarks screen becomes "Bookmarks & notes" and lists each book's
  highlights and notes in reading order.
• Import a whole folder: "Import folder…" on the Books screen imports
  every EPUB under it, Calibre's Author/Title/ layout included, with
  progress, a Stop button and a summary of what was imported, skipped
  or failed.
• Windows support: Pitaka builds and runs on Windows, and CI checks it
  there.

Changed

• The app is named "Pitaka" (capitalised) in its window title, the
  Start Menu and the installed apps list. The library stays where it
  was.
• Chapter search loads its embedding model at a pinned version, so a
  change to the published model can't silently mix with chunks indexed
  by the old one.

Fixed

• Malformed XHTML no longer loses the rest of a chapter: broken markup
  is skipped and parsing carries on after it. Text using HTML entities
  such as   or —, or a bare & ("Faith & Reason"), is
  kept instead of dropped, and a bare & in a book's title no longer
  fails the import. Books already in the library need re-importing to
  pick this up.

Downloads: Windows: Pitaka_0.3.0_x64-setup.exe. Linux: the .AppImage, or the .deb/.rpm for your distro. The Windows installers aren't code-signed yet, so Windows SmartScreen will warn before running them; choose More info → Run anyway. SHA256SUMS lists each file's checksum.

v0.2.0

Added

• Find chapters by meaning, in a build with --features semantic:
  Search gains a Passages / Chapters switch, and a chapter result opens
  the reader at the paragraph that matched. A small embedding model
  runs locally, downloaded once on first use. Books are indexed in the
  background when imported, with progress on the Books screen; books
  already in the library need re-importing to be included.

Changed

• A chapter with no <h1>/<h2> takes its name from the book's table
  of contents before falling back to the page's <title>, so books
  that style headings as <div>s no longer show the book's name for
  every chapter. Books already in the library need re-importing to
  pick this up.
• The project page shows chapter search, says the browser demo doesn't
  have it, and puts the app icon beside the name.

There are no prebuilt binaries yet; see Install to build from source.

v0.1.0

The first tagged version, covering everything built so far.

Added

• Import EPUBs into a local SQLite library, skipping a book whose file
  is already in it from any path.
• Full-text search across every book, ranked, with the matched words
  highlighted. "learn" also finds "learning" unless Exact words is
  ticked; case and diacritics are ignored; quoted phrases, prefix*
  and AND/OR/NOT work.
• Search terms expand to their curated transliteration variants, so
  dharma also finds "dhamma".
• A continuous-scroll reader with a chapter sidebar. Opening a search
  result centres and flashes the matching paragraph.
• Bookmark paragraphs into named folders.
• Remove a book from the library, so it can be re-imported.
• A launch screen with separate Books, Bookmarks and Search screens.
• The Paper visual style: bundled serif fonts that cover Pali
  diacritics, and a dark mode that follows the system theme.
• A project page on GitHub Pages with a browser demo of the real
  interface, using one CC0 book.
• Contributing guide, security policy, code of conduct, and issue and
  pull request templates.

Changed

• Chapter titles come from the first <h1>/<h2>, falling back to the
  page's <title>.
• Books that use <div> rather than <p> for paragraphs are imported,
  each content block becomes one paragraph, and text split across
  inline tags (such as drop caps) is joined without adding spaces.
• The EPUB 3 navigation document is no longer imported as a chapter.
• The app shows an error and quits if it can't open its library,
  instead of crashing with nothing on screen.

Fixed

• Punctuation in a search no longer breaks the query.

Security

• Search snippets are rendered as text rather than HTML, and the app
  sets a Content Security Policy, so a book's text can't inject markup
  or script into the app.

There are no prebuilt binaries yet; see Install to build from source.